Healthcare providers spend many hours each week writing clinical notes to document patient visits.
Kwindla Hultman Kramer, co-founder and CEO of Daily, developer of WebRTC, an open-source tool that allows developers to work with video and audio, has worked with healthcare professionals and telehealth application engineers to develop a technology capability with the aim of reduce documentation time by 80% or more.
Daily’s new APIs enable seamless integration of SOAP and other clinical note workflows into virtual care platforms. Daily’s WebRTC developer platform offers a range of HIPAA compliant APIs and supports Safari and iOS for one-click, no-download telehealth video calling.
We sat down with Kramer to discuss the role of AI in telehealth, documentation during virtual care visits, and the new documentation API.
Q. Why do you think telemedicine is a fertile ground for the application of artificial intelligence?
A. From a technology perspective, an important aspect of telehealth interactions is that all audio is already captured digitally, ready for transcription and summary. This makes telemedicine a good starting point for adding new AI tools to healthcare workflows.
The category of AI tools currently receiving the most attention and adoption are called large language models. For example, ChatGPT is a user interface on top of OpenAI’s major language models, GPT-3.5 and GPT-4.
The latest, advanced LLMs are quite good at taking unstructured text as input and producing structured data as output. This is something computers have never been capable of before, and this new capability is a big reason technologists are so excited about these new tools.
The transcript of a telehealth visit is a good example of “unstructured data” that has a lot of valuable information embedded in it. Until recently, the only way to understand a telemedicine transcript was for someone to read it and then extract some of the information for entry into a medical record system.
For example, now large language models can automatically create clinical documentation in specific formats.
The use of telemedicine grew rapidly during the COVID-19 pandemic, accelerating regulatory, billing, and technology changes that had already begun. As telemedicine has now become an established part of healthcare, there is an opportunity to think about ways in which adding AI tools to telemedicine – thoughtfully and in a non-disruptive way – can help increase access to care, save physicians significant amounts of time and improve healthcare. patient results.
Q. What can AI do for providers’ SOAP and other clinical notes that providers cannot do themselves?
A. In general, I don’t think we should try to find things AI can do that healthcare providers can’t do themselves. I think we should try to find things that AI can do to save time and allow healthcare providers to focus on working with patients.
No one is drawn to healthcare as a career because they want to spend time writing clinical documentation. SOAP notes and such are very important. But taking SOAP notes isn’t exactly creative work and doesn’t make much use of a healthcare provider’s unique expertise and human touch, especially when compared to interacting with patients.
An AI tool that can automatically produce a first draft of a SOAP note – for approval or editing by the healthcare provider – has the potential to save enormous amounts of time. Potentially ten hours or more per week. We want this note to be as similar as possible to what the caregiver would make if he or she were to write it down.
In the long term, it may be valuable to investigate what new AI tools can do differently or better than humans. I don’t think we should just dismiss this line of research.
But at this point we’re starting to feel pretty confident that today’s new AI tools can do a new class of relatively routine things that take a lot of time, with perhaps minimal loss of quality. If that turns out to be true, it means healthcare providers can spend more time with their patients, or more time with their families, or both. That’s a big problem.
Q. Your company just released the AI-Powered Clinical Notes API for telehealth. Who is this API for and how do they integrate it with virtual care technology? Furthermore, what are the results that the API should deliver?
A. Our customers are software developers who build telehealth applications. We specialize in the video and audio components of software development. Our APIs – application programming interfaces – are designed to be used as part of a complete product or service.
As we’ve grown, we’ve expanded our features to include things bordering on video and audio, like transcription, analytics, and now AI tools.
Our goal in everything we do is to enable the creation of new software that has value in the world. We don’t just work in healthcare, but healthcare is an especially important part of what we do because the value of helping people get the best medical care possible is so clear and so motivating.
In the case of the new AI-powered Clinical Notes APIs we just released, we started working on this because our customers told us that their clients – healthcare providers – typically spend ten hours or more per week writing clinical documentation and our asking if we had any ideas on how we could reduce that burden.
We had some ideas, so we worked closely with some of our customers to test, evaluate and iterate. Getting even a first public version of this stuff right is obviously important in several ways: data privacy, quality of output, reliability, and reliability of service.
Q. How else do you see artificial intelligence being applied to telemedicine in the future?
A. Some of the things we’re seeing that seem likely to have an impact in the near term include making it easier for both patients and providers to prepare for visits, creating more detailed analytics of care outcomes, and different approaches of real-time ‘copilots’. ” that provide access to information during a session.
Further afield and more experimentally, I am interested in the progress being made in ‘multimodal’ AI models. These are large models trained on a huge amount of data, including text, images and audio, and sometimes data from temperature, inertial and other specialized sensors.
There is potential to use all the data our digital devices collect about us to help doctors diagnose and treat patients. There are obvious privacy issues, but overall I think we can do a good job of regulating access to and privacy of healthcare data (much more than we do for other types of data).
We can perhaps think of this as an extension of telemedicine to include passive and “always on” health monitoring and diagnostics.
Follow Bill’s HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.